Presentation on theme: "PRECIS Providing REgional Climates for Impact Studies The Hadley Centre regional climate modelling system A presentation to the UNFCCC workshop on non-Annex."— Presentation transcript:
1 PRECIS Providing REgional Climates for Impact Studies The Hadley Centre regional climate modelling systemA presentation to the UNFCCC workshop onnon-Annex 1 national communication preparation guidelinesbyRichard Jones, Met Office Hadley Centre, UK8 April 2003
2 Outline Background: Motivation and PRECIS overview PRECIS components and their applicationFuture climate scenarios available from PRECISApplying PRECIS in impacts assessmentLimitations of PRECIS and future developments
3 MotivationUNFCCC requirement to assess national vulnerability and plans for adaptationUNFCCC requirement to submit National CommunicationsBoth need estimate of impactsImpacts need detailed scenarios of future climateScenarios are best produced locally, using expert knowledge and increasing ownershipRegional Climate Models (RCM) provide climate information with useful local detail including realistic extreme events. Developing countries are the most vulnerable to climate change. Hence the need for detailed climate change scenarios to assess their national vulnerability. The development of a PC version of the Hadley Centre’s RCM is addressing this need by allowing developing countries to generate their own national scenarios of climate change for use in impact studies. This will lead to technology transfer and capacity building within developing countries.
4 Predicting impactsTo predict future climate change, we first need projections of emissions of greenhouse gases and other constituents. These emission scenarios have been developed in the IPCC Special Report on Emission Scenarios (SRES) and reflect a number of different ways in which the world might develop. The concentrations of these gases is calculated using carbon cycle and chemistry models taking as input the above emission scenarios. The concentration scenarios are used as input into the Global coupled climate models to compute global climate projections. The current resolution of the atmospheric part of a typical GCM is about 250 km in the horizontal, and of the ocean is 125 to 250 km. This resolution is not high enough to represent the fine-scale detail that characterises the climate in many regions of the world. Also, it will be insufficient for the requirements of most impacts models. Hence there is the need to add regional detail and here the technique shown is via regional climate models (as this is what PRECIS uses) though there are other regionalization techniques.Finally the impacts models require climate scenarios as inputs. The climate scenario will be constructed by combining the climate change prediction (from the RCM) with a description of the current climate as represented by the observational data (the observed “baseline” climate) or by using RCM predictions of current climate as the baseline and of future climate as the scenario..
5 Who is PRECIS for?Government scientists from less developed countries involved in vulnerability and adaptation studiesA regional model’s domain usually encompasses several countries, so it is hoped that neighbouring countries will collaborate to produce ensemblesIt is hoped to draw on local resources, such as expertise in local climatePRECIS is mainly aimed at government scientists from developing countries involved in vulnerability and adaptation studies, thus in general with either a background in meteorology/climate or climate impacts. As RCM domains generally include several countries it is hoped that PRECIS would be co-operatively used by neighbouring countries. Involving scientists within the region in the use of PRECIS is expected to increase the value of resulting climate scenarios by using their expertise to provide appropriate assessment and interpretation of the scenarios.
6 Who funds PRECIS?The new regional climate modelling system, PRECIS, has been developed at the Hadley Centre and is sponsored by the DEFRA, DFID and UNDP:The UK Department for Environment, Food and Rural Affairs (DEFRA) funds the development of the Hadley Centre’s RCM and specific collaborations focused on its use over India, southern Africa and China.The UK Department for International Development (DFID) funds the PC version of the Hadley Centre’s RCM.The United Nations Development Programme’s (UNDP) National Communications Support Unit (NCSU) has provided funds from the Global Environment Facility (GEF) to support the provisions of training materials relevant to PRECIS for experts in developing countries.
7 Hardware requirements PC running under the Linux operating systemMemory : 512MB minimum; 768MB recommendedMinimum 60GB disk space + offline storage for archiving dataSimulation speed proportional to chip speedHow fast does it go?30 year integration, 100x100 grid pointsT3E (supercomputer) months (36 Processors)PC (Intel P4 2.8 GHz) monthsThe hardware specification provided here is a minimum for convenient operation of PRECIS though less online disk space would be acceptable if offline tape storage was available for storing of PRECIS output and PRECIS boundary conditions. (This would require experiments to be run in shorter segments, e.g. 1-2 years at a time). In terms of processor speed, it is advisable to obtain the fastest processor available.
8 PRECIS: PC system components DomainScenarioPeriodDiagnosticsThis schematic introduces the components of the PRECIS software package.Run PRECIS
9 PRECIS components: full list User-interface to set up RCM experimentsThe latest Hadley Centre RCMData-processing and graphics softwareBoundary conditions from latest Hadley Centre GCMTraining course and materialsPRECIS website and helpdeskIn addition to the main software components, the first three items listed here, there are three other important components. The first is the set of boundary conditions which is needed to run the PRECIS RCM and is derived from an archive of global data at the Hadley Centre. The second is this training course and associated materials which provide the background to make most appropriate and best use PRECIS. The third is the PRECIS website which is a source of these and other materials, hosts an interface to request boundary data and provides a forum for requesting advice and sharing experiences with PRECIS.
10 What is a Regional Climate Model? Comprehensive physical high resolution climate model that covers a limited area of the globe, usually including the atmosphere and land surface components of the climate system, and containing representations of the important processes within the climate system (e.g., cloud, radiation, rainfall, soil hydrology).A Regional Climate Model (RCM) is a high resolution climate model that covers a limited area of the globe, typically 5,000 km x 5,000 km. RCMs are based on physical laws represented by mathematical equations that are solved using a three-dimensional grid. The typical horizontal resolution of an RCM is 50 km. Hence RCMs are comprehensive physical models, usually including the atmosphere and land surface components of the climate system, and containing representations of the important processes within the climate system (e.g., cloud, radiation, rainfall, soil hydrology). Many of these physical processes take place on much smaller spatial scales than the model grid and cannot be modelled and resolved explicitly. Their effects are taken into account using parametrizations by which the process is represented by relationships between the area or time averaged effect of such sub-grid scale process and the large scale flow.
11 Regional Climate Models (RCMs) Limited area models driven at the boundaries by GCM observed boundary data.Resolution of 50kmThe nested regional climate modelling technique consists of using initial conditions, time-dependent lateral meteorological conditions and surface boundary conditions to drive high-resolution RCMs. The driving data is derived from GCMs (or analyses of observations) and can include GHG and aerosol forcing. A variation of this technique is to also force the large scale component of the RCM solution throughout the entire domain.To date, this technique has been used only in one-way mode, i.e. with no feedback from the RCM simulation to the driving GCM. The basic strategy is thus to use the global model to simulate the response of the global circulation to large scale forcings and the RCM to a) account for sub-GCM grid scale forcings (e.g. complex topographical features and land cover inhomogeneity) in a physically-based way; and b) enhance the simulation of atmospheric circulations and climatic variables at fine spatial scales.
12 Boundary conditions Requests through PRECIS website helpdesk The boundary conditions for the PRECIS RCM are clearly an integral part of the system but as they comprise a very substantial amount of data (20-30 Gbytes for a 30-year simulation) they have to be supplied separately. They are stored online at the Hadley Centre and will be made available on request through a web-based interface. The data will then be supplied on a storage medium specified by the user.
13 Training course and supporting material Training in the use of PRECIS will focus on:Background science including uncertaintiesInterpretation of PRECIS results by regional expertsConstruction of regional climate change scenariosBuilding capacity in countries/regions using PRECISPRECIS will be supplied with:a workbook covering the background science, system description and the uses and limitations of PRECISa technical manual explaining technical details about the system and how install and to use itTraining workshops will be held to explain to users of PRECIS relevant scientific background so that they understand the components of PRECIS and how to make best use of it. It will explain the potential usefulness and also the limitations of the data PRECIS can provide and how they can be used to construct climate change scenarios. It will also give examples of using PRECIS data in impacts studies. An important aspect of the workshop will be obtaining feedback from users on what they are hoping to use PRECIS for and to encourage and facilitate collaboration between users of PRECIS.As reference material for users, much of the content of the workshop will be supplied to users (in addition to the workshop presentations) in a workbook, covering the scientific aspects, and a technical manual, covering the installation and use of PRECIS.
14 Support and follow-up Helpdesk email, phone web-based discussion WebsitenewsupdatesdatasetsresourcesCollaboration/workshopsPRECIS users will be able to get assistance from Hadley Centre staff by or phone and also via a web-based discussion group. The webiste will provide information on what PRECIS is being used for, future plans, any updates to PRECIS, relevant datasets which users may find useful (small enough to be downloaded) and a list of resources, such as relevant papers and reports. Finally, Hadley Centre staff will be keen to collaborate on research and scientific publications involving PRECIS and to attend, and possibly help organise, workshops where collaborators are presenting and discussing results from PRECIS.
15 Applications of RCMs Regional detail when simulating current climate Realistic detail in climate change predictionsResolution of islands, smaller countries e.g. ItalyRealistic simulation of extreme eventsProvides comprehensive data for impact modelsSome example applications of RCMs are presented in the following slides demonstrating the features listed here.
16 WINTER PRECIPITATION OVER BRITIAN Observed, and simulated with RCM and GCM Most land areas have mountains, coastlines etc on scales of 100km or less and RCMs can take account of the effects of much smaller-scale terrain than GCMs. The example here shows simulated and observed precipitation over Great Britain. The observations clearly show enhanced rainfall over the mountains of the western part of the country, particularly the north-west, which is well captured by the RCM but completely absent in the GCM.
17 SIMULATION OF A TROPICAL CYCLONE in the PRECIS regional climate model Global climate model Regional climate modelThe coarse resolution of GCMs does not allow them to properly represent smaller scale weather features such as tropical cyclones and these are much better simulated at higher resolution in RCMs. This example shows the surface pressure patterns from one day of a GCM simulation and that from the corresponding RCM simulation where the latter has developed a cyclone in the Mozambique Channel.
18 WINTER DAILY RAINFALL OVER THE SW CAPE, SOUTH AFRICA SW Cape winter rainfall: Observed - red, GCM - green, RCM - blueBetter detail in simulating daily events is also seen when examining the frequency of these. Here, the frequency of winter daily rainfall for the SW Cape region of South Africa over a range of intensities in thirty year simulations of current climate are compared with observations. For low and medium intensity events both the GCM and RCM agree well with the observations but for the more intense and extreme events the GCM severely underestimates their frequencies whereas the RCM is much more realistic.< >40mm/day
19 CHANGES IN MONSOON PRECIPITATION between the present day and the middle of the 21st century Global climate model Regional climate modelThe extra detail seen in RCM simulations of current climate is also seen in RCM predictions of climate change. In many regions, an aspect of climate that is predicted to change is the circulation, i.e. wind directions and strength. This, for example, will modulate the strength of flow incident on mountains and the position of windward rainy areas and those downwind in a rain-shadow. This effect will only be seen, however, if mountains are properly resolved in the model and explains the substantial differences seen in the patterns of predicted change in monsoon rainfall in the GCM and corresponding RCM simulations shown here. The RCM resolves the Western Ghats which consequently receive much of the higher rainfall which the GCM distributes over the interior much of which is, realistically, a rain-shadow area in the RCM which, in fact, predicts reductions.
20 SUMMER TEMPERATURES in the 2080s compared to the present day, due to A2 emissions Global climate model Regional climate modelThe coarse resolution of the GCM means that many islands or smaller countries (e.g. Italy) are not resolved and are represented as sea. As the land has a much lower thermal inertia than the sea when it is heated it warms much faster. The results of this can be seen in this example where RCM predictions of summer temperature increases over land areas not resolved by the GCM are much higher, often by a factor of two. These predictions are clearly more realistic.Climate on islands changes very differently to the surrounding Mediterranean Sea, and can only be predicted using an RCM
21 Future climate scenarios available from PRECIS PRECIS can provide:climate scenarios for any regionan estimate of uncertainty due different emissionsan estimate of uncertainty due to climate variabilityData available from PRECIScomprehensive for atmosphere and land-surfacegrid-scale box average quantitiesmaximum time resolution one hourPRECIS has boundary conditions available for the whole globe and thus can run over any region. They are also available from a series of GCM experiments which have been run using different emissions scenarios and, for one emissions scenario, run from different initial conditions thus providing an estimate of the influence of natural variability.The data available from PRECIS includes a very large number of variables describing both the state of the atmosphere and the land-surface. These represent grid-box average, and not point, values and have a maximum time resolution (for a smaller range of variables) of one hour.
22 CO2 EMISSIONS PROFILES under IPCC SRES scenarios Emission scenarios are plausible representation of future development of emissions of substances that are potentially radiatively active (e.g., greenhouses and aerosols), based on a coherent and internally consistent set of assumptions about driving forces (such a demographic and socio-economic development, technological change) and their key relationships. The SRES scenarios set comprises four scenario families: A1, A2, B1, B2. The scenarios within each family follow the same storyline. The A1 family includes three groups reflecting a consistent variation of the storyline (A1T, A1FI and A1B). Hence the SRES scenarios consist of six distinct scenario groups, all of which are plausible and together capture the range of uncertainties associated with the driving forcing and emissions.The SRES scenarios do not include additional climate initiatives, which means that no scenarios are included that explicitly assume implementation of the United Nations Framework Convention on Climate Change or the emissions targets of the Kyoto Protocol. In particular, none involves a stabilisation of concentrations of greenhouse gases.Anthropogenic emissions of CO2 are shown for the six illustrative SRES scenarios, A1B, A1FI, A1T, A2, B1 and B2. For comparison the previous emissions for the IPCC business as usual scenarios, IS92a, are also shown.Source:IPCC
23 GLOBAL TEMPERATURE RISE due to four SRES emissions scenariosThis figure shows the response of global mean temperature to the four emissions scenarios with the extra lines in yellow and red showing the influence of natural variability on the evolution. In these cases, simulations were started using different initial conditions but the same emissions data.
24 Applying PRECIS in impacts assessments Obtain current state from impacts model via:a) observed data e.g. as used in model validationb) RCM control dataObtain future state from impacts model using:RCM climate changes (in means, variability, etc) applied to observed data (compare with a)RCM future climate data directly (compare with b)Assessing impacts of climate change are usually done by feeding in baseline current climate data and then data for a future climate state into an appropriate model. The baseline data can be observed values, in which case the future state is obtained by perturbing these baseline values using climate change data derived from the RCM. Alternatively, the baseline data could come from an RCM simulation of the current climate with the future state then simply being data from the RCM simulation of the future climate.
25 BAY OF BENGAL CYCLONE simulated by the PRECIS regional climate model Resulting storm surge simulated using the Proudman Oceanographic Laboratory modelThis example shows where RCM data has been fed directly into the impacts model, in this case of storm surges over the northern Bay of Bengal.
26 CHANGE IN ANNUAL SURFACE RUNOFF between today and the 2080s, for A2 emissions In this example, the impact of climate change on surface runoff over southern Africa is calculated from a GCM and RCM using changes in mean climate from the models applied to an observed baseline.Southampton University hydrological model
27 Current limitations of PRECIS 10 year simulations take a month or moreIt is a model hence imperfect, i.e. uncertainty due to use of only one RCM is not capturedUncertainty due to imperfect driving data (i.e. from only one GCM) is not capturedResolution of 50km or 25km insufficient for some impacts applicationsCoastal or marine impacts have to be inferred (e.g. storm surges)The speed of PRECIS is limited by the speed of the PC processor which currently means that 10 year simulations take a month or more to complete. As with any model, it is imperfect, and a model to provide high resolution climate information it is only one from one class of techniques to provide this information. In considering other aspects of uncertainty in future climate scenarios, it currently only using boundary conditions from one GCM. Also, even at a resolution of 25km, it will not resolve all the details required for many impacts applications. Finally, the PRECIS RCM currently only represents the atmosphere and land surface so coastal and marine impacts of climate change have to be inferred or are not represented at all.
28 PRECIS: treatment of uncertainties PRECIS accounts for uncertainties due to:Future emissions - uses 4 SRES scenariosNatural variability - uses initial condition ensemblesPRECIS does not account for uncertainties due to:The conversion of emissions to concentrations - carbon cycle or chemistry (except sulphur cycle) not representedUncertainty in the response of the climate system - only boundary conditions from 1 GCM are usedUncertainty in the detailed climate change - only one RCM is usedUncertainties about future climate arise from a number of different sources. Depending on the climate scenario construction method, some of these uncertainties will be explicitly represented in the resulting scenarios, while others will be ignored.Of the five key sources of uncertainties relating to climate scenario construction:Alternative emissions futuresUncertainty in converting emissions to concentrationsUncertainties in converting concentrations to radiative forcingUncertainties in modelling the climate responseUncertainties in converting model response into inputs for impact studiesonly the first, the sulphur cycle aspect of the second and the climate variability aspect of the fourth are accounted for by PRECIS.
29 CHANGE IN S ASIA SUMMER RAINFALL with PREDICTED BY NINE GCMs (A2 emissions) The nine climate change predictions from the models assessed in the IPCC Third Assessment Report (TAR) for summer precipitation over the Indian subcontinent indicate the level of uncertainty resulting from the imperfect representation of the climate system in the coupled models. Here, all models predict an increase in rainfall, though in a range of 8-24%.A2 emissions scenario
30 % CHANGE IN WINTER PRECIPITATION This figure is similar to the previous one though looking at the pattenrs of change predicted by the models, this time over the UK in winter. This shows the large range of patterns of predicted change from the 9 models. It also including results from three ensemble members of the Hadley Centre GCM (marked UKMO) which were initialised with different starting conditions. These are thus three realisations of the climate of the 2080s consistent with the A2 emissions scenarios and demonstrate the uncertainty coming from the internal variability of the climate system and that it is smaller than that from the inter-model ranage.
31 REPRESENTATION OF THE PHILIPPINES WITH DIFFERENT MODEL RESOLUTIONS This figure shows the benefits of increasing resolution from 50 to 25km, which is part of the PRECIS functionality, though in many applications, e.g. for catchment-scale hydrology, this is still not sufficient.25km RCM resolution km RCM resolution GCM 300km resolution
32 FUTURE DEVELOPMENTS Continuously upgraded to new processors Ability to run PRECIS RCM from different GCMsat least two to be added this yearrelevant cooperation from GCM centres being soughtIntercomparisons of all RCMs being organisedNo plans for PRECIS to incorporate other RCMsResolution of model being improved to 10kmRCM with ocean component being builtPRECIS is continuously being tested on new processors to ensure it works so enabling the quickest run times to be achieved.The ability to run PRECIS using boundary conditions from different GCMs will be added as these come available. Currently, an interface to the ECHAM GCM is being built and another to the CSIRO GCM will follow and discussions are happening with other global modelling centres to obtain access to the necessary data.There is a version of the PRECIS RCM which has been run at 25km and this functionality is included in PRECIS for short integrations or small areas. The Hadley Centre is currently building a new RCM which will be able to be run at resolutions of up to 10km and will include a regional ocean model.
33 Final remarks You can contact us via: Tel:Fax:Web-site for information on PRECIS